public/repo2markdown/src/util, branch master[no description]http://www.cgit.farzat.xyz/public/repo2markdown/atom/src/util?h=master2026-06-10T17:04:12ZMove outer_backticks to a util module2026-06-10T17:04:12ZA Farzata@farzat.xyz2026-06-10T17:04:12Zurn:sha1:f11361ee359c1eaccc36559a2fa5b1b98da10ba8
This is because it shall be used by other functions later on.
It was made more generic by accepting raw bytes instead of strictly
UTF-8 strings. This allows future functions to call it without having to
convert the contents to UTF-8 strings first.
Group similar languages together in testing2026-06-09T16:55:36ZA Farzata@farzat.xyz2026-06-09T16:55:36Zurn:sha1:4d787ed64fef6bb3c1559ce4ee5663d4723a2558Support YAML, JS, and TOML2026-06-08T01:05:14ZA Farzata@farzat.xyz2026-06-08T01:05:14Zurn:sha1:091ae954417b7895062691e3713ebf406c986f29Prioritize detection through extensions2026-06-07T18:15:45ZA Farzata@farzat.xyz2026-06-07T18:15:45Zurn:sha1:a088bbbbd0eed36b9b4eb3d39dc9bc306d33f185
While shebangs take priority on Unix-like systems, the shebang detection
here is very basic and not guaranteed to be correct, so its use should
be secondary.
Detect sh as bash2026-06-07T18:12:50ZA Farzata@farzat.xyz2026-06-07T18:12:50Zurn:sha1:b3fe9879da5386df853c4c5d583eb040341195da
This is based on the assumption that LLMs recognize bash language fences
more as they are trained more on them. Another assumption is that LLMs
wouldn't care much about the difference unless instructed.
Ensure shebangs properly start with #!2026-06-07T18:04:37ZA Farzata@farzat.xyz2026-06-07T18:04:37Zurn:sha1:133dba2bc968c9081144887ceb19bd9711c16df6Refactor language detection in a util module2026-06-07T17:58:37ZA Farzata@farzat.xyz2026-06-07T17:53:20Zurn:sha1:99d21463574bb88fa86f176c2e4ad154be5e1353Move render_filename to utils as display_path2026-06-06T00:39:39ZA Farzata@farzat.xyz2026-06-06T00:22:07Zurn:sha1:e26d589dc129faa3018546239fb2ac4675f5753c